- Edited
I brought another GTX 970 yesterday (Same as the 1st MSI Gaming OC) and I SLI ; Working perfect still Crysis 3 is the most demanding game (graphically to me) on 4K - 3840 X 2160 using DSR /Ultra Setting and 2x TXAA I am tipping 55-60 Fps :) ; Sure Call of Duty -Advanced Ware Fare on, Middle Earth: Shadow of Mordor , Evil With In , Wolfenstein: The New Order are all 4K /Ultra ; linking it to my 40inch Led TV through HDMI with 4K DSR the real power of graphics kick out
To Remove Any Bottle Necks I am Using:
-I7 4790 @ 3.6 GHZ
-16 GB Ram 1666 MHZ
-Z97 Chipset
-One Side 140mm Fan for Direct VGA Cooling
-Games are Installed on SSD - 512 GB Crucial 100MX
-PSU: Thermal Take Tough Power Fully Modular - 80 Plus Gold Certified / SLI Certified
-The System Runs Very Stable ; with No Heat Issues (I am monitoring the GPUs Temperatures / Voltages) via MSI After Burn 4.0 ; the Smart Cooling of MSI Twin Frazor 5.0 is really great combined with the Heat / power efficiency of the Maxwell GM 204 which is great for SLI much better compared to AMDs R9s R8s and R9s and Nvidia Kepler on the 7 Series (Tried All).
I can say I am future for couple of year (2-3 years) - Just Saying :) since the issue now is the Vram as more games being released playing on 4ks with Ultra Textures are tipping the 4 GB Ceiling on some games (even in 1440p) ; as such Nvidia will release soon the 8GB Vram version of the GTX 970 and 980s Plus the 1080 and 1070 will be released in the future with 8GB Vram by Default as reports
being lately emerged started to indicate.
Let us wait and see if First off, we should see the R9 390X being the first GPU to be built on TSMC's new 20nm manufacturing process, but the card is rumored to arrive with High Bandwidth Memory, or HBM. HBM is 3D stacked memory technology that will offer an incredible amount of bandwidth on the already-fast GDDR5 tech that is used, with around 100% more bandwidth, all while using less power which might remove this bottle neck -This new memory technology is AMD's & Hynix's answer to supplying the next generation of APUs & GPUs with the necessary memory bandwidth to reach maximum performance, essentially eliminating one of the major bottlenecks of GPU compute present today.
HBM stands for High Bandwidth Memory, and it's made essentially by stacking multiple memory chips on top of each other via die-stacking technology, creating a 3D memory structure made up of a large number of memory floors. ; but I am an Nvidia loyal - The Volta GPU will introduce stacked DRAM (dynamic random access memory which I think they getting the Tech from Nvidia this time lool - competitors game since AMD will release HBM Compatible GPUs in the 1st Quarter of 2015 in R9390 before Nvidia Release on the next Volta Architecture in 4th Quarter of 2014 ). On a technical level, this is accomplished by having multiple layers of DRAM stacked atop each other on the same silica substrate, then drilling a hole through the silicon and connecting each layer.
1TB/s of memory bandwidth means you could move the contents of a jam-packed Blu-ray disc through the chip in a mere 1/50th of a second. Consider that NVIDIA’s Titan GPU — its latest high-end graphics card — can handle about 288GB/s, and you can only imagine the positive implications for a wide range of people depending on graphics processing horsepower, from game developers to graphic artists
Important Note:
-HBM doesn't increase bandwidth by raising memory frequency similar to GDDR5 rather by significantly increasing the number of channels through which the data can travel. Similarly HBM is slower than GDDR5 in terms of frequency but can move orders of magnitude more data in the same time period as GDDR5
To Remove Any Bottle Necks I am Using:
-I7 4790 @ 3.6 GHZ
-16 GB Ram 1666 MHZ
-Z97 Chipset
-One Side 140mm Fan for Direct VGA Cooling
-Games are Installed on SSD - 512 GB Crucial 100MX
-PSU: Thermal Take Tough Power Fully Modular - 80 Plus Gold Certified / SLI Certified
-The System Runs Very Stable ; with No Heat Issues (I am monitoring the GPUs Temperatures / Voltages) via MSI After Burn 4.0 ; the Smart Cooling of MSI Twin Frazor 5.0 is really great combined with the Heat / power efficiency of the Maxwell GM 204 which is great for SLI much better compared to AMDs R9s R8s and R9s and Nvidia Kepler on the 7 Series (Tried All).
I can say I am future for couple of year (2-3 years) - Just Saying :) since the issue now is the Vram as more games being released playing on 4ks with Ultra Textures are tipping the 4 GB Ceiling on some games (even in 1440p) ; as such Nvidia will release soon the 8GB Vram version of the GTX 970 and 980s Plus the 1080 and 1070 will be released in the future with 8GB Vram by Default as reports
being lately emerged started to indicate.
Let us wait and see if First off, we should see the R9 390X being the first GPU to be built on TSMC's new 20nm manufacturing process, but the card is rumored to arrive with High Bandwidth Memory, or HBM. HBM is 3D stacked memory technology that will offer an incredible amount of bandwidth on the already-fast GDDR5 tech that is used, with around 100% more bandwidth, all while using less power which might remove this bottle neck -This new memory technology is AMD's & Hynix's answer to supplying the next generation of APUs & GPUs with the necessary memory bandwidth to reach maximum performance, essentially eliminating one of the major bottlenecks of GPU compute present today.
HBM stands for High Bandwidth Memory, and it's made essentially by stacking multiple memory chips on top of each other via die-stacking technology, creating a 3D memory structure made up of a large number of memory floors. ; but I am an Nvidia loyal - The Volta GPU will introduce stacked DRAM (dynamic random access memory which I think they getting the Tech from Nvidia this time lool - competitors game since AMD will release HBM Compatible GPUs in the 1st Quarter of 2015 in R9390 before Nvidia Release on the next Volta Architecture in 4th Quarter of 2014 ). On a technical level, this is accomplished by having multiple layers of DRAM stacked atop each other on the same silica substrate, then drilling a hole through the silicon and connecting each layer.
1TB/s of memory bandwidth means you could move the contents of a jam-packed Blu-ray disc through the chip in a mere 1/50th of a second. Consider that NVIDIA’s Titan GPU — its latest high-end graphics card — can handle about 288GB/s, and you can only imagine the positive implications for a wide range of people depending on graphics processing horsepower, from game developers to graphic artists
Important Note:
-HBM doesn't increase bandwidth by raising memory frequency similar to GDDR5 rather by significantly increasing the number of channels through which the data can travel. Similarly HBM is slower than GDDR5 in terms of frequency but can move orders of magnitude more data in the same time period as GDDR5